Content distribution networks are distributed systems of computing devices deployed across a global communications network to distribute content. Content distribution networks may include servers, storage, and associated components that distribute content including web objects e.g., text, graphics, and scripts, downloadable objects, e.g., media files, software, and documents, applications, e.g., e-commerce and portals, live streaming media, on-demand media, and social networks. Content distribution networks may be deployed in many geographic locations often over multiple backbones. Content distribution networks may include any number of service nodes in any number of varying geographic locations depending on architecture. Content distribution networks may direct user requests for content to service nodes that are optimal in some manner, e.g., service nodes that may provide content with the fewest hops, service nodes with the least number of network seconds away from the requesting client device, or service nodes with the highest availability in terms of performance. When optimizing for cost, content distribution networks may chose service nodes that are least expensive. In many instances, content distribution networks may optimize service node selection for both performance and cost.
Often, clients are routed to their closest service node using inaccurate or incomplete information, which provides acceptable level of accuracy for latency tolerant applications. For very low latency applications, however, a more accurate means of routing clients to the closest service node is required.